对等网络环境下基于多属性约束的任务调度算法研究  

Research on Task Scheduling Algorithm Based on Multi-attribute Constrains in the Environment of P2P Network

作　　者：陈晶[1] 陈端君[2] 

机构地区：[1]燕山大学信息科学与工程学院计算机系,河北秦皇岛066004 [2]北京交通大学海滨学院,河北黄骅061100

出　　处：《小型微型计算机系统》2012年第12期2571-2574,共4页Journal of Chinese Computer Systems

基　　金：河北省自然科学基金项目(F2011203092)资助

摘　　要：为了提高对等网络环境中的任务调度效率,提出将资源属性与任务属性相结合的多约束任务调度算法MDSA.在该算法的实现过程中,依据被调度的任务类型和隶属函数的定义,将对等网络环境划分为不同的虚拟社群,并将不同类型的任务提交到对应需求的虚拟社群中;将对等网络抽象描述为资源属性的集合,并定义了待调度的任务集合;以资源和任务属性的多约束信息为基础,利用待调度任务的时限属性信息形成调度队列,借助于预判机制和二级调度的思想对任务进行调度,并应用NS2和GridSim工具包对MDSA调度算法进行了测试.通过对仿真结果中的数据进行分析和比较,验证了多属性约束调度算法MDSA的优势.To improve the task scheduling efficiency of peer-to-peer network environment, task scheduling algorithm MDSA （integra- ting Multi-resource attribute with Deadline Scheduling Algorithm } is proposed. In the implementation process of the algorithm, divid- ing P2P network environment into different virtual communities based on the attribute of the membership function, and submitting dif- ferent type tasks to correlative needs virtual communities. Peer to peer network is abstracted as a collection of resource attribute, and collection of tasks is defined. On the basis of multi-attribute constraints, producing the scheduling queue according to deadline and lo- cation of the node information, scheduling task by means of two level scheduling ideas and prediction mechanism. MDSA scheduling algorithm was tested by applying NS2 and GridSim toolkit. The simulation results show that algorithm MDSA is superior by compa- ring and analyzing the result results.

关 键 词：多属性约束 虚拟社群 二级调度 任务时限 资源定位 隶属度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]